Fix some conversion issues

README.md

@@ -47,16 +47,18 @@ First, clone the repo:
 ## Linux
 
 - Install system requirements
-  - `cat install/apt-requirements.txt | xargs sudo apt-get install -y`
   - Optional: Install tesseract 5 by following [these instructions](https://notesalexp.org/tesseract-ocr/html/) or running `install/tesseract_5_install.sh`.
   - Install ghostscript > 9.55 by following [these instructions](https://ghostscript.readthedocs.io/en/latest/Install.html) or running `install/ghostscript_install.sh`.
+  - Install other requirements with `cat install/apt-requirements.txt | xargs sudo apt-get install -y`
 - Set the tesseract data folder path
   - Find the tesseract data folder `tessdata` with `find / -name tessdata`.  Make sure to use the one corresponding to the right tesseract version if you have multiple!
   - Create a `local.env` file in the root `marker` folder with `TESSDATA_PREFIX=/path/to/tessdata` inside it
 - Install python requirements
   - `poetry install`
   - `poetry shell` to activate your poetry venv
+- Update pytorch as needed since poetry doesn't play nicely with it
   - GPU only: run `pip install torch` to install other torch dependencies.
+  - CPU only: Uninstall torch, then follow the [CPU install](https://pytorch.org/) instructions.
 
 ## Mac
 

convert.py

@@ -92,14 +92,14 @@ if __name__ == "__main__":
         storage=settings.RAY_CACHE_PATH,
         _temp_dir=settings.RAY_CACHE_PATH,
         dashboard_host=settings.RAY_DASHBOARD_HOST,
-        log_to_driver=False
+        log_to_driver=settings.DEBUG
     )
 
     model_lst = load_all_models()
-    model_refs = [ray.put(m) if m else None for m in model_lst]
+    model_refs = ray.put(model_lst)
 
     # Dynamically set GPU allocation per task based on GPU ram
-    gpu_frac = settings.INFERENCE_RAM // settings.VRAM_PER_TASK if settings.CUDA else 0
+    gpu_frac = settings.VRAM_PER_TASK / settings.INFERENCE_RAM if settings.CUDA else 0
 
     print(f"Converting {len(files_to_convert)} pdfs in chunk {args.chunk_idx + 1}/{args.num_chunks} with {total_processes} processes, and storing in {out_folder}")
     futures = [

install/apt-requirements.txt

@@ -2,6 +2,7 @@ tesseract-ocr
 libtesseract-dev
 libmagic1
 ocrmypdf
+tesseract-ocr-eng
 tesseract-ocr-deu
 tesseract-ocr-por
 tesseract-ocr-spa

marker/settings.py

@@ -12,6 +12,9 @@ class Settings(BaseSettings):
     TORCH_DEVICE: str = "cpu"
     INFERENCE_RAM: int = 40 # How much VRAM each GPU has (in GB).
     VRAM_PER_TASK: float = 2.5 # How much VRAM to allocate per task (in GB)
+    DEBUG: bool = False # Enable debug logging
+    DEFAULT_LANG: str = "English" # Default language we assume files to be in, should be one of the keys in TESSERACT_LANGUAGES
+
     SUPPORTED_FILETYPES: Dict = {
         "application/pdf": "pdf",
         "application/epub+zip": "epub",
@@ -19,7 +22,6 @@ class Settings(BaseSettings):
         "application/vnd.ms-xpsdocument": "xps",
         "application/x-fictionbook+xml": "fb2"
     }
-    DEFAULT_LANG: str = "English" # Default language we assume files to be in, should be one of the keys in TESSERACT_LANGUAGES
 
     # PyMuPDF
     TEXT_FLAGS: int = pymupdf.TEXTFLAGS_DICT & ~pymupdf.TEXT_PRESERVE_LIGATURES & ~pymupdf.TEXT_PRESERVE_IMAGES
@@ -48,7 +50,7 @@ class Settings(BaseSettings):
     OCR_ALL_PAGES: bool = False # Run OCR on every page even if text can be extracted
     OCR_PARALLEL_WORKERS: int = 4 # How many CPU workers to use for OCR
 
-    # Nougat Model
+    # Nougat model
     NOUGAT_MODEL_MAX: int = 512 # Max inference length for nougat
     NOUGAT_TOKEN_BUFFER: int = 256 # Number of tokens to buffer above max for nougat
     NOUGAT_HALLUCINATION_WORDS: List[str] = ["[MISSING_PAGE_POST]", "## References\n", "**Figure Captions**\n", "Footnote",
@@ -57,7 +59,7 @@ class Settings(BaseSettings):
     NOUGAT_MODEL_NAME: str = "0.1.0-small" # Name of the model to use
     NOUGAT_BATCH_SIZE: int = 4 if TORCH_DEVICE == "cuda" else 1 # Batch size for nougat, don't batch on cpu
 
-    # Layout Model
+    # Layout model
     BAD_SPAN_TYPES: List[str] = ["Caption", "Footnote", "Page-footer", "Page-header", "Picture"]
     LAYOUT_MODEL_MAX: int = 512
     LAYOUT_CHUNK_OVERLAP: int = 64